Відео

Langfuse is opensource and can be selfhosted. Or you can go old reliable route and log everything to use something like ELK stack on your logs

Mu gut feeling tells me, that the only how you can actually have a long term memory without flaws and scaffolding, which always accompany RAG, is continuous fine tuning.

Great video, thank you! Btw, which LLM was used as a judge model.. and doesn't using an LLM model over the other as a judge bias the evaluation

How can we store the fitted model? I want to use the fitted BM25 model repeatedly on my app. Is there a way to keep it?

Bro, that chewing gum …. Just dont

Thanks for the video, Can you pls make a video on LLM deployment in AWS/LLMOps

how to config Jina to scrape what we want in a specific website? (image, avatar, links, etc)

7:25 transforming HTML to Markdown ? its amazing. Why don't we do this to the entire Web and get rid of all the front-end code in Javascript and all the bullshit and leave just the Text so we humans can read without the gibberish. Someone needs to make a Browser that uses LLMs to read the stupid HTML and get rid of the crap and present a beautiful markdown that can be directly rendered using DirectX12 to my quad-hd screen. Maybe I could make that. How about we make the web go back to HATEOAS

Hmmm. CHATGpt recommended your video🎉. Just thought I'd share incase you aren't aware.

Can someone just give me the conclusion

Can anybody explain to me why I should use paid AI services for web scraping rather than using my own scripts with open source libraries like Selenium?

Thank you for the video, Can't wait if you can do the same for TTS & TTS please 🙏

Horrible disrespectful chewing. Could not follow it due to chewing.

Interesting videos on your channel but could you please not chew gum as the mic picks it up too much? Keep it up 👍

DELETE THIS VIDEO RN, ITS TO POWERFUL 🫨

Excellent video!

greatttt

Would really appreciate a similar video for translation.

Your videos are fun to watch

Great Video! The chewing is a Little bit annoying though.😂

maybe i've missed it on the video, but why do you think Exa was not able to answer any of the queries correctly? like 0% every time

This is awesome stuff, very high value investment of 15 minutes.

great

That’s a great tutorial/ code example. Thanks a lot! It inspired me to include something similar in my project. I especially like the idea to first read out the sitemap and then loop through the pages! And thanks for the other comments with ideas about more scraping tools, I need to look them up! At the moment I’m using DataForSEO, which works pretty well, too.

Great work! Keep it up!

awesome vid!

Pure gold, thanks for sharing

Always very good content…..

More content here! app.catswithbats.com/90d4bd29

Its a cool product but only issue is Jina getting blocked as a bot, so its not making it past the "Are you human?" screen.

This video made my SaaS possible thanks - I had no idea 5 months ago what LLM scraping was.

why this video isn't in trending list !!!!!!!!!!!!!!!

Hey i wanna connect with you guys , is there a online community kind of thing?

Always happy when you upload new video

i watch these videos, i have evolved

allergies season is finally over!

you're a great teacher!

Who else was disappointed after clicking the thumbnail and seeing a dude?

You still need a big computing and storage to actually build her.

Nice demo <3

Hey, great content! Thanks for sharing your knowledge. However, instead of just using tsvector in PostgreSQL, you can leverage sparse vector search by utilizing the pg_search extension, right?

that sounds illegal?

crappy camera work, good content

video lacks a good hook

Haha, in the old days, I just used lynx URL | perl -e '....' to achieve somehow the same, nowadays lynx usually doesn't work anymore just as every website doesn't contain content any more, but executables, and of course Perl became the most hated language around, so nobody likes to parse anything with perl anymore, my colleagues very often even struggle with regexes, so year, using 3rd party tool to get content out of websites and then ask an LLM to answer your question probably is a million or billion times or worse compute inefficient than the old way, but can be done by more people. But I remember we analyzed huge product catalogues like amazon, icecat, .... to get product description of every category with our approach for probably one some dollars in total while processing Gigabytes of raw content (so close to a billion tokens in these terms).

thank you for the presentation. What is the best (cheapest) way to scrape data from instagram? all I need travers graph of accounts to find accounts with certain threshold of followers?

Are you having difficulty breathing?

Firecrawl is not as good as scrapfly ... firecrawl seems like it was invented by a single guy in his bedroom while scrapfly is enterprise grade.

firecrawl seems cool but I still feel like jina might be a bit better overall...

i like to use scrapy because its open source and LLM friendly